We compared global DNA methylation alterations between normal and testicular tumor cells using methylated DNA immunoprecipitation and tiling array hybridization (MeDIP-chip). A high resolution cytosine methylation map of the human germ cell cancer was obtained and more than 6 thousand Differential Methylated Regions (DMRs) between normal and TGCT cells were identified. More than 70% of DMRs resided in intergenic regions. Promoter methylation accounted for 9%. About 1/3 (27%) of these genes demonstrated the perceived relationship between promoter methylation and gene expression, i.e. hypermethylation associated with suppression of gene expression. [unreadable] [unreadable] A focal analysis of DMRs located in the regulatory regions of annotated genes yielded 207 differentially methylated genes. We selected 3 candidate genes for further characterization in primary tumor tissues. There three genes were Apolipoprotein L domain containing 1 (APOLD1), Retrotransposon gag domain containing 1(RGAG1) and Protocadherin 10 (PCDH10). The open reading frame of APOLD1 encodes an apolipoprotein-L domain-containing protein whose function is unknown. Remarkably, APOLD1 is located in 12p13.1, a TGCT susceptibility locus identified previously by genetic linkage analysis. RGAG1 is an X-linked retrotransposon-derived neogene with unknown function. Expressed sequence tags (EST) of RGAG1 were found predominantly in testis, suggesting that this retrogene might be important in male germ cell development. PCDH10 encodes a membrane protein for cell adhesion. It had been implicated as a tumor suppressor gene in nasopharyngeal, esophageal, breast, colorectal, cervical, lung and, hepatocellular carcinomas. Differential methylation of these three genes observed in cultured tumor cells was confirmed in primary testicular tumor tissues by bisulfite sequencing and methylation sensitive PCR. Differential expression of these genes in tumor and control tissues was confirmed by RT-PCR. Thus, cultured cells could be used as a model for studying the mechanism of altered methylation in tumors and these genes may serve as novel non-invasive epigenetic markerer for molecular diagnosis of TGCT. [unreadable] [unreadable] The role of these intergenic DMRs is not clear. They could simply be a consequence of inappropriate epigenetic establishment during PGC differentiation, or might have a hidden regulatory role such as the maintenance of genomic stability or chromatin condensation. Another possible function of non-genic DMRs is the regulation of noncoding RNAs. We mapped the non-genic DMRs to current noncoding RNA database and found that 3 miRNAs (miR-199a-2, miR-124a-2 and miR-184) were hypermethylated in TGCT. Notably, miR-124a was first identified in embryonic stem cells, suggesting that this miRNA may be important in differentiation. Coincidentally, this miRNA was identified to be epigenetically silenced in colon cancer and consequently activate an oncogene CDK6. Furthermore, we also found that 3 snoRNAs (hb11-240, aca22 and aca8) were hypomethylated. Epigenetic changes of noncoding RNAs might lead to deregulation of genetic networks in a wider spectrum, as a single miRNA is capable of regulating numerous target genes.